Discover more from ChinaTalk
Endless Frontier Neutered!
"The Endless Frontier Act (EFA) is one of the most important legislative proposals you’ve never heard of — or at least it was."
Yesterday I published an Emergency ChinaTalk podcast with Sam Hammond of the Niskanen Center on the butchering of the Endless Frontier Act over the past week. It’s one thing to not like the Tech Directorate (more context in last week’s edition, this was initially going to get $100bn over 5 years and now is a rounding error in this bill), but the Senate Science committee voting 23-6 to shrink the total size of the R&D commitment for the entire bill it’s truly puzzling. Lujan’s amendment dramatically reshaped the form of the legislation and seems like it was drawn up in a really slapdash manner.
What we have now are single-digit percentage boosts to the NSF’s funding, some money out of left field for the Department of Energy, and a plus-up for National Labs like Los Alamos (thanks to the National Labs Caucus which, perhaps unsurprisingly, seems more focused on pushing money to their states’ arguably duplicative labs than addressing domestic innovation challenges). None of this stuff is wasteful spending, of course, but the dialling down of this bill’s ambition is pretty depressing.
Also, mainstream media didn’t follow the amendment process and is getting this story totally wrong.
But I guess we’ll get some money for fabs which is nice I guess?
See below for a reprint of Sam Hammond’s recent piece for his take on the issue.
How Congress Ruined the Endless Frontier Act
The Endless Frontier Act (EFA) is one of the most important legislative proposals you’ve never heard of — or at least it was.
First introduced last year by Senators Chuck Schumer and Todd Young, the bill would have established a new Technology Directorate at the National Science Foundation (NSF) with a DARPA-like program structure equipped with flexible hiring and grant-making authorities. With a $100 billion budget over five years, the Directorate would have been empowered to use grants, contracts, prizes, and cooperative agreements with industry, academia and research institutes to push the frontiers of U.S. innovation in ten broad areas, ranging from cutting-edge technologies like Artificial Intelligence and quantum computing, to more mature but no less important sectors like robotics, manufacturing, biotechnology, advanced energy technology and material sciences.
That was then, this is now. While the mission of the Technology Directorate remains nominally intact, the scale of the EFA’s ambitions were suddenly curtailed late last week. Following a flurry of Senate Commerce Committee amendments, the $100 billion originally envisioned for the Directorate dropped to just $29 billion over five years, with $54.9 billion reassigned to the NSF-proper. Of that $29 billion, only 15% or $4.35 billion is reserved for the Directorate’s core R&D mission, with the rest earmarked for scholarships, testbeds, academic tech transfer, and investments in regional “innovation centers.”
But at least the NSF is getting a big boost, right? Wrong. The $54.9 billion to the NSF replaces their existing funding rather than being a supplement. That leaves only $12.9 billion in new NSF funding, of which $8.23 billion is tied to promoting STEM education. In short, what was sold as a ~$100 billion boost in federal support for R&D is now less than $40 billion in new spending, of which less than $10 billion is reserved for anything resembling research or development.
The Endless Logrolling Act
Federal support for research and development has some of the highest benefit-cost ratios of anything the government funds. And as a form of long-term public investment, funding for R&D is the least of even the most ardent fiscal conservative’s concerns. When one considers that the state of Louisiana alone spends a billion dollars a year on industrial tax abatements, or that the SBA is about to begin awarding $16 billion in grants to entertainment venues affected by the pandemic, the decision to short-change the Endless Frontier Act is simply incomprehensible.
How did this happen? The short answer is that the U.S. Senate Commerce Committee is beyond dysfunctional. Some level of congressional logrolling (the practice of adding amendments and exchanging favors for votes) is to be expected, but in this case logrolling completely eclipsed discussion of the underlying bill. When Senator Schatz of Hawaii introduced an amendment to crack down on the sale of shark fins, for instance, a reporter on the scene tweeted, “Not gonna lie, did not expect an extended debate on shark finning in this Endless Frontier markup. But we’re now discussing shark steaks vs. shark-fin soup.”
With chum in the water, the feeding frenzy had only just begun. Next was Senator Cantwell, the committee’s chairwoman. As the Senator from Boeing Washington, Cantwell forced an amendment adding $10 billion to NASA’s budget in order to give Blue Origin a shot at winning the lunar landing contract NASA has already awarded to SpaceX at a fraction of the cost, all while requiring NASA to continue developing an unnecessary launch system that Boeing just happens to have the contract for.
Senator Luján of New Mexico took the biggest chunk out of the Endless Frontier Act by far, however, with an amendment that reallocated the bulk of its budget to the Department of Energy’s National Labs. The influence of the NSF’s own advocacy was apparent as well, given worries that the EFA would give the Technology Directorate too much autonomy and detract from the agency’s traditional focus on inquiry-oriented science in favor of applied research and technology. By the end of the amendment process, what is now known as the U.S. Innovation and Competition Act ballooned to over 1420 pages long. The merits of these and other provisions are beyond the scope of this commentary.
Applied vs. Basic Science is a False Dichotomy
This last concern — that the EFA’s Technology Directorate would be too autonomous and too focused on technological applications — is particularly misguided. Indeed, in many ways the Directorate’s autonomy and applied focus is the entire point.
Securing technological leadership, whether vis-à-vis China or for technological progress’s own sake, requires more than just support for basic science. As technology writer Dan Wang has noted,
Both the design process and production process generate useful information, and dislocation makes it difficult for that information to circulate. … The U.S. has little position in making high-end precision manufacturing equipment. When it comes to factory automation systems, machine tools, robot arms, and other types of production machinery, the most advanced suppliers are in Japan, Germany, and Switzerland. I think the reason that the U.S. has little position can be tied directly to the departure of firms from so many segments of manufacturing. How do engineers work on the design of automation systems if they don’t have exposure to industrial processes?
When it comes to technological development, scientific breakthroughs, while essential, are just the tip of the spear. Moving research from the lab and into production is important not just from a technology-transfer perspective, but also for accelerating the technology cycle itself.
The dramatic decline in the cost of solar panels over the past decade is a case in point. As the technologist Ramez Naam never tires of pointing out, “Solar has plunged in price not through innovation in panel technology, but by innovation that reduces the cost of *manufacturing* the panels.”
Solar companies spend less than 1% of their revenue on R&D. Exponential cost reductions have instead come through a learning-by-doing process, as those same companies scale-up their production. This isn’t a story of pure scale economies, however. Rather, the need to scale forces genuine process innovations; things like a production engineer realizing, through hands-on experience, that they could improve efficiency 10% by tweaking this or that chemical solvent. Thus the notion that we first do basic science first and then translate those findings into applied technology isn’t just wrong — it’s often just the reverse. The solar panel production boom has even inspired some scientists to talk of “Solar-Driven Chemistry,” as insights derived from the solar industry’s learning-by-doing continue to spill-over into new ideas for basic research.
As an added bonus, supporting domestic technology ecosystems helps build a moat around American intellectual property by default. Consider that Boeing could give China the design specs for the 787 and all its associated patents, and China would be scarcely closer to manufacturing a wide-body airliner. Indeed, they almost certainly have those designs and patents in their possession already, but without Boeing’s decades of cumulative manufacturing “know-how,” China’s joint-venture with Russia is struggling to get their 787 competitor off the ground, much less turning a profit.
China’s government understands this even if America’s lawmakers do not, and is thus working day and night to build-up their industrial know-how. As the China analyst Emily De La Bruyere recently explained to Jordan Schneider for the ChinaTalk podcast, the Chinese government spends very little on basic science because they’re confident they can always steal the most important discoveries. Their R&D spending is thus instead overwhelmingly in technology and applications:
The argument is that what matters today is the applications of science and technology––the sort of networks you build with it. For example, your ability to deploy telecommunications … it’s not that you got the patent; it’s that you’ve got its application internationally. To do that, what matters is capturing scale and being able to build and deploy. If that’s what you’re going for, it’s okay to have a slight lag in when you get the patent and when you get the really cutting edge, as long as you can apply it to scale to the most people efficiently.
The Chinese orientation appears to be focusing on that rather than on basic R&D, which creates this tremendous asymmetry vis-à-vis the U.S. and really vis-à-vis the entire global system because there’s just a different competition underway. And that absolutely changes how the U.S. can or should respond to the extent that this is a scientific and technological contest because it’s not a matter of just pouring resources into basic research: it’s about competing for applications.
In the context of the recent changes to the EFA, De La Bruyere’s bottom-line should make you want to pull your hair out:
The U.S. says there is this contest and we need to invest in science and technology because we’re competing against China. But what they don’t ask is “how does China compete” or “what are our resources actually going to fuel, now that we’re benchmarking against a competitor” or “are they in fact just going to fuel our competitor?”
Why “Embedded Autonomy” Matters
The second edge China has over the United States isn’t so much technological as institutional. While often characterized as a command-and-control style economy, the day-to-day of Chinese industrial policy is surprisingly decentralized. Five year plans like “Made in China 2025” mostly serve to set high-level targets and aspirations, helping to coordinate the expectations of bureaucrats and industry partners at multiple levels of government.
A similar story holds true for the successful examples of industrial policy in Korea, Japan, and the United States. As Steven Vogel argues in Level Up America: The Case for Industrial Policy and How to Do it Right, investments in technology and industrial capacity work best when done through institutions with “embedded autonomy.” The central government should set clear, outcome-oriented goals with mechanisms to evaluate progress, but leave implementation and execution to mission-driven organizations with the autonomy to take risks and act nimbly. Embedded autonomy is particularly important when the agency in question has “a strategic position as the central nodes in networks of collaboration among industrial sectors and firms,” and might thus be vulnerable to special interest capture. “An industrial policy driven by bold public missions accompanied with deliberate communication strategies,” writes Vogel, “would be less vulnerable to capture and more amenable to effective implementation.”
What would an institution with mission-driven “embedded autonomy” look like in the context of U.S. technology policy? Well, a lot like the Endless Frontier Act, at least as originally conceived. Its Technology Directorate would be insulated from special interest politics, and have clear, mission-driven research goals with impartial mechanisms for evaluating progress and periodically updating its core research areas. Flexible hiring authority would enable the Directorate to recruit highly capable program directors; the sorts of personnel that can be entrusted to implement their respective missions with minimal bureaucratic red-tape. The authority to award grants, prizes and contracts to industry would support applied research and tangible production beyond a narrow focus on basic science. And in serving as a node between industry and academia, the Directorate’s autonomy would help limit the risk of capture.
The new version of the Endless Frontier Act does the opposite on essentially every point. The Directorate’s budget is gutted, with only $4.35 billion left for its core research mission. Rather than focus on scaling and applications, diverting money to the National Labs, colleges and the NSF-proper reinforces our increasingly sclerotic approach to basic science. Rather than have autonomy, the Directorate’s governance is subordinated to the NSF Director and the Energy Secretary — honorable people, just with their own distinct priorities. And rather than be free from special interest capture, its authorizing legislation builds-in capture by academia and industry incumbents from the get go.
American Science Needs Institutional Innovation
If China overtakes the United States technologically or geopolitically, it won’t be because we lacked the financial resources or awarded too few STEM degrees. It will be because of our utter lack of institutional innovation, driven by incumbent lock-in and our own leaders’ failures of imagination.
To wit: The National Defense Research Committee was created in 1940, and despite its short existence, helped launch research into some of the most important technologies the United States used to win World War II, from radar and sonar to what later became the Manhattan Project. Its creation was owed to Vannevar Bush, the father of the NSF and the man who first dubbed science the “endless frontier” (sound familiar?). Bush managed to get a meeting with President Roosevelt on June 12, 1940, and brought a single sheet of paper describing the proposed agency, which Roosevelt approved in ten minutes.
That kind of institutional speed and innovation from the U.S. government is unthinkable today, and yet Bush got it done despite fierce incumbent opposition. As Bush recounts in his autobiography,
There were those who protested that the action of setting up NDRC was an end run, a grab by which a small company of scientists and engineers, acting outside established channels, got hold of the authority and money for the program of developing new weapons. That, in fact, is exactly what it was.
Replace “weapons” with advanced manufacturing and robotics, next generation hardware and software, applied biotechnology and materials science, and much, much more, and that’s in some sense what the Endless Frontier Act is, too. Or should I say, was?